Ethical Decision-Making: Frameworks, AI Ethics, and Professional Codes

TAKEAWAYS:

• Ethical decision-making follows structured frameworks (e.g., moral reasoning models) to balance rights, duties, and consequences.
• AI ethics requires addressing bias, transparency, and accountability (e.g., Google’s AI hiring tool discriminating against women).
• Professional codes (IEEE, ACM) provide guidelines for IT professionals to navigate conflicts like whistleblowing or data misuse.
• Real-world examples (e.g., Kathmandu traffic optimization via AI) show how ethical choices impact society and technology.
• Licensing and education ensure professionals adhere to ethical standards (e.g., NTC’s cybersecurity certifications for IT workers).
• Workplace ethics involves monitoring, SLAs, and privacy (e.g., Daraz’s data handling policies for customer trust).

Core Concepts: Ethics and Ethical Theories

Ethics is the study of right and wrong behavior, while ethical theories provide frameworks to evaluate decisions. Below are the key theories relevant to IT professionals:

1. Major Ethical Theories

mindmap
  root((Ethical Theories))
    Utilitarianism["Utilitarianism: Greatest good for the greatest number"]
      Example["AI in healthcare: Balancing patient data privacy vs. life-saving diagnostics"]
    Deontology["Deontology: Duty-based ethics (Kant’s Categorical Imperative)"]
      Example["Refusing to hack a system even if it helps a friend"]
    Virtue Ethics["Virtue Ethics: Moral character (e.g., honesty, integrity)"]
      Example["A programmer reporting a security flaw in eSewa’s payment system"]
    Rights-Based Ethics["Rights-Based: Protecting individual rights (e.g., GDPR)"]
      Example["Ncell’s obligation to notify users of data breaches"]

Worked Example: AI in Loan Approvals

• Scenario: A bank uses AI to approve loans. The algorithm rejects 80% of applicants from rural areas due to biased training data.
• Ethical Analysis:
  • Utilitarian: Denying loans harms rural economies but may reduce fraud.
  • Deontological: The bank has a duty to fairness (Kant’s ethics).
  • Virtue-Based: An ethical AI developer would audit the algorithm for bias.
  • Rights-Based: Applicants have a right to non-discriminatory services (GDPR/PDPA).

Ethical Decision-Making Frameworks

Professionals use structured models to resolve ethical dilemmas. Below is a 4-step framework (adapted from IEEE/ACM):

Example: Whistleblowing at a Tech Firm

• Problem: An employee discovers Pathao’s drivers are being paid below minimum wage due to algorithmic manipulation.
• Steps:
  1. Identify: Unfair pay practices.
  2. Facts: Internal documents show driver earnings data.
  3. Stakeholders: Drivers, Pathao management, government labor laws.
  4. Alternatives:
     • Report internally (risk of retaliation).
     • Leak to media (violates NDAs).
     • File a complaint with NTC (legal route).
  5. Ethical Theories:
     • Deontology: Duty to expose injustice.
     • Utilitarian: Helps drivers but may harm Pathao’s reputation.
  6. Decision: File anonymously with NTC (balances duty and consequence).

AI Ethics: Challenges and Case Studies

AI introduces unique ethical challenges, such as bias, accountability, and transparency. Below are real-world examples:

1. Bias in AI Systems

pie
  title AI Bias Sources
  "Training Data Bias" : 45
  "Algorithmic Design" : 30
  "Lack of Diversity in Teams" : 25

Case Study: Google’s AI Hiring Tool

• Issue: Google’s AI tool penalized women’s resumes by associating keywords like "women’s" with lower rankings.
• Ethical Violation:
  • Deontological: Discriminated against a protected class.
  • Utilitarian: Reduced diversity in tech teams.
• Solution: Google audited the algorithm and retrained it with unbiased data.

2. Accountability in Autonomous Systems

• Example: A self-driving car (e.g., Tesla) causes an accident. Who is liable?
  • Manufacturer (design flaws)?
  • Software Developer (coding errors)?
  • User (misuse)?
• Ethical Framework: Distributed accountability (shared responsibility among stakeholders).

Professional Codes of Ethics

IT professionals follow codes from organizations like IEEE and ACM to guide conduct. Below is a comparison:

Worked Example: Data Privacy at eSewa

• Scenario: eSewa collects user transaction data but sells it to third parties without consent.
• Ethical Violation:
  • ACM Code: Violates privacy rights.
  • NTC Guidelines: Breaches data protection laws.
• Action: An ethical IT professional would:
  1. Report internally (if no response).
  2. File a complaint with NTC.
  3. Publicly disclose (last resort, per whistleblower protections).

In the Real World

1. eSewa’s Ethical Dilemma

   • Idea Used: Data privacy and consent (GDPR/PDPA compliance).
   • How: eSewa must obtain explicit user consent before sharing data. Ethical IT staff audit systems to ensure compliance.

2. AI in Kathmandu Traffic Management

   • Idea Used: Utilitarian ethics (maximizing public good).
   • How: AI optimizes traffic lights to reduce congestion. An ethical decision would prioritize pedestrian safety over faster vehicle flow.

3. Ncell’s Ethical AI for Customer Service

   • Idea Used: Transparency and fairness.
   • How: Ncell’s AI chatbot must disclose when it’s human vs. AI and avoid discriminatory responses (e.g., rejecting calls from rural areas).

Workplace Ethics: Monitoring, SLAs, and Privacy

IT professionals must balance productivity monitoring with employee privacy. Below are key considerations:

1. Employee Monitoring Policies

classDiagram
  class Employee {
    +Work Hours
    +Internet Usage
    +Keyboard Activity
  }
  class Employer {
    +Performance Metrics
    +Security Compliance
    +Legal Limits
  }
  Employee -->|"Monitored For"| Employer : Productivity
  Employer -->|"Must Respect"| Employee : Privacy Rights

Example: Remote Work Monitoring at Daraz

• Policy: Daraz tracks employee screen time but cannot access personal chats.
• Ethical Conflict:
  • Utilitarian: Improves productivity.
  • Rights-Based: Employees have a right to privacy (Nepal Labor Act).

2. Service Level Agreements (SLAs)

SLAs define ethical obligations between IT service providers and clients. Example:

• NTC’s SLA for Internet Providers:
  • Uptime Guarantee: 99.9% (ethical to avoid false promises).
  • Data Security: Encrypt user traffic (deontological duty).

Exam Tip

1. Define Clearly:

   • Always start with definitions (e.g., "Ethical decision-making is the process of evaluating alternatives using ethical theories...").
   • Marks are lost for vague answers.

2. Use Real-World Examples:

   • Examiners love Nepali case studies (e.g., eSewa, Ncell, Daraz).
   • Link theories to AI, data privacy, or workplace ethics.

3. Framework Questions:

   • For "explain ethical decision-making," use the 4-step model (Identify → Facts → Stakeholders → Alternatives).
   • For AI ethics, discuss bias, accountability, and transparency.

4. Compare Codes of Ethics:

   • Tables (like the IEEE vs. ACM comparison) score full marks for structured answers.

5. Avoid Common Mistakes:

   • ❌ "Ethics is just following laws." → Wrong: Ethics goes beyond laws (e.g., whistleblowing).
   • ❌ Ignoring stakeholders in case studies. → Always list users, companies, and regulators.